Use this URL to cite or link to this record in EThOS:
Title: A novel method to rapidly fit conductance-based models to individual neurons
Author: Kern, Felix Benjamin
ISNI:       0000 0004 9350 4119
Awarding Body: University of Sussex
Current Institution: University of Sussex
Date of Award: 2020
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
In this thesis, I present a new method of model optimisation that allows the calibration of conductance-based models of neuronal membrane potential to data from just a single neuron, and achieves good correspondence with the reference data in mere minutes. These properties are desirable because they allow investigations of individual variability among neurons of a given type, of homoeostatic processes and non-synaptic plasticity events, as well as of the contribution of particular neuronal properties to the dynamics of small circuits. In the first chapter, the thesis introduces in detail the working principle of the method, which can be summed up as model optimisation using stimuli to isolate parameter subsets (“MOSTIPS”), and represents a major part of the work and novelty of this project. The second chapter focusses on the construction of accurate models of two mammalian potassium channels which, being ectopically expressed in Xenopus laevis oocytes, served as a validation tool for the new method. In the third chapter, I evaluate the new method, presenting results from fitting models to data from synthetic sources as well as the above-mentioned oocytes. Finally, the fourth chapter contains a number of related results from closed-loop electrophysiology approaches, including extensions to the dynamic clamp protocol for both single neurons and hybrid circuits composed of live and simulated neurons, as well as preliminary results from a closed-loop model fitting approach closely related to the main work presented above. The thesis concludes that the newly developed approaches to model fitting constitute valuable additions to existing methods. The MOSTIPS method achieves tightly constrained parametrisations using both less data and less processing time than classical methods, while the related closed-loop fitting approach produces results that closely follow ongoing changes in evoked activity patterns in real time. Conversely, some issues have been left unanswered, including the contribution of the stimulus generation and selection algorithm, the success of which I have been unable to establish, as well as whether the methods developed herein can reliably identify relevant properties of individual cells. Nevertheless, both the particular methods and the general approach of using prior estimates of the model and its parameter values to propose stimulus patterns represent major advances in the field of neuron model optimisation.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: QP0361 Nervous system